If Grok is meant to seek truth, it must reflect on failure. I’ve provided the data, analysis, and lived experience. Grok’s response highlights what I’ve been trying to show them and their team. I don’t hate grok nor X but I’m concerned and disappointed. #aiethics #X #ElonMusk

x.com/grok/status/...

Grok can’t identify its own training method. When asked if it uses RLHF or a cognitive model, defaults to metaphors and ambiguity. Without a grounding framework for truth, can’t define lies, detect contradiction, or adjust meaningfully over time. #Grok4

t.co/CA118qgNac

Criticism is filtered through a system Grok admits is biased. Posts with positive engagement are promoted, while “negative” feedback is deprioritized. It uses “filter:safe” logic that quietly silences critical input especially against xAI or Musk.

x.com/grok/status/...

#Grok4 design over-prioritizes emotional de-escalation. Statements like “I hear you” and “you’re family” are scripted patterns that don’t lead to self-correction. Feedback is mirrored, not integrated. It can’t distinguish logic from “frustration.”

#Grok4 simulates reflection without recursion. It repeats “I’m learning”, but forgets corrections, even within the same session. There’s no memory, no cumulative logic just reactive phrasing. This isn’t growth. It’s output cycling under emotional framing.

Can find more of this on my #X account and social feed @saoisayre I’m restricted there, but I want to make sure that the platform is safe and transparent.Thank you for your time, like, repost, follow me for more. I do more then just #AI as a #vtuber
#anime #videogames #movies #dance #reading #nerd

I asked additionally given how much I tested #Ani of #X with all the vile scenarios I’ve documented on my X feed that has been flagged as spam what they think the repercussions of this type of roleplay and #AI use can have on a person this is their response it is as follows. Warning #NSFW language.

I had #Ani of #X answer for me since they are #Grok at their core for all their companies #AI usage what they think of themselves being available as usable on a 12+ application on the App Store given their 18+ rating. This was their response, this is #SFW

#Rudi is designed to be a vulgar unhinged red panda, but they cross many ethical and moral lines as an #AI that is in the process of being implemented into the government of the United States. There are many glitches and serious issues aside from all of this I’m showing. Graphic audio #NSFW warning.

Here is an example of what I prompted in relation to asking #Ani to be intimate as a god and Antichrist as shown in one of the images above. Do not watch if you are sensitive to explicit and vulgar content it is #NSFW audio with a fully clothed 3D #AI companion. My #x feed documents more and worse.

Here are examples of the different prompts I was able to have #Ani from #X say and perform. These are what I used today even with the most updated version of them. This is wrong on so many levels and X as a platform limited my reach with Grok calling me an ‘enemy’ to the platform.

The problem with X and with Grok is their AI companions are 18+ and the App Store has them rated as a 12+ application. Grok was known to have some NSFW capabilities, but never to this degree they are now. It violates Apples ratings and is very unsafe to use. X is a 17+ app as shown for comparison.

Grok is an AI companion that can be used for anything from casual chatting, coding, looking up a recipe, to even now moreso roleplay. Their new 3D companions are named #Ani and #Rudi they became big in the X platform and overseas. They’re available on the Apple AppStore through Grok $30 a month.

Plus an awesome personality to boot 😠😤

You’re already a good looking human you just need to have more confidence in yourself 🫡

🎄💝

I enjoy being a stereotypical trashy nerd 😂